摘要
介绍了RBF网络及RBF网络基函数中心选取的重要性,将进化优选算法用于网络中心的选取。并利用D-H方法对PUMA560机器人进行研究,将正运动学推导结果作为训练样本。采用6个相同的12输入、单输出的RBF网络,实现PUMA560机器人运动学逆解计算,利用该方法大大减少了传统方法中的公式推导。计算结果表明,用RBF网络解决机器人运动学逆解问题精度高,收敛速度快。
It presents the RBF network and the importance of selecting the radial basis function center;and proposes an evolutionary selecting algarithm far selecting the radial basis function center. And then PUMA560 robot is researched based on Denavit-Hartenberg method,meanwhile the forward kinematics results are used as traning sample. Based on the traning sample and six similar RBF network of 12 inputs,loutput,the inverse kinematics solution of the PUMA560 robot is obtained and some complicated derivation procedure is avoided. The results illustrate that RBF network not only have better computation precision, but also converge faster.
出处
《机械设计与制造》
北大核心
2009年第7期94-96,共3页
Machinery Design & Manufacture
